Please proceed at first to the Welcome Students pages here:
http://hci-kdd.org/scientific-working-for-students

SP.64 Building a Digital Pathology Deep Learning Machine
Status
open
Worktype
All types of student work (Projekt Informationssysteme, Bachelor, Master, PhD - please look up [3])
Keywords
Deep Learning Machine, Digital Pathology
Goals
For our Digital Pathology project we will carry out some experiments with TensorFlow and compare this with fast a own built deep learning machine. More information in a personal conversation. Read the story by Lukas Biewald from February, 1, 2017 on building a super fast deep learning machine for under $1,000 [1] before making contact - this is a cool project for nerds 🙂 Read also [2].
For Students of
Students with interest in doing experimental work on a cool topic (AI + Medicine)
Additional Information

[1] https://www.oreilly.com/learning/build-a-super-fast-deep-learning-machine-for-under-1000

[2] https://hci-kdd.org/2017/10/26/digital-pathology-worlds-fastest-scanner-now-working-graz

[3] 706.505 Projekt Informationssysteme, 706.800, 706.801 Bachlorarbeit Informatik, 706.804, 706.805, Bachelorarbeit Softwarenetwicklung-Wirtschaft, 706.902 Bachelorarbeit Information and Computer Engineering, 706.501 Master-Projekt, 706.502 Master Project Web and Data Science, and 706.046 AK HCI https://online.tugraz.at/tug_online/wblvangebot.wbshowlvoffer?ppersonnr=5313

 

SP.63 Making Deep Learning Transparent (MADEL)
Status
open
Worktype
All types of student work (Projekt Informationssysteme, Bachelor, Master, PhD - please look up [5])
Keywords
explainable AI, Visualization of Deep Learning, Explainability
Goals
In this project the advantages, disadvantages and the practicability of the application of using deconvolutional networks [1] for making convolutional networks [2] transparent shall be experimentally explored. The contribution is relevant for our MAKEpatho project [3]. The tests shall be performed with real-world images produced by our local facilities [4]. More details are given in a personal conversation, but before making contact please read [1] and check if this would interest you.
For Students of
Students with interest in machine learning/artifical intelligence basic research
Required Knowledge
Understanding of convolutional networks and interest
Additional Information

[1] Matthew D. Zeiler & Rob Fergus 2014. Visualizing and understanding convolutional networks. In: D., Fleet, T., Pajdla, B., Schiele & T., Tuytelaars (eds.) ECCV, Lecture Notes in Computer Science LNCS 8689. Cham: Springer, pp. 818-833, doi:10.1007/978-3-319-10590-1_53. [and refer to arXiV:1311.2901 https://arxiv.org/abs/1311.2901   (2,691 citations as of 23.11.2017

[2] Alex Krizhevsky, Ilya Sutskever & Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. In: Pereira, Fernando, Burges, Christopher J.C., Bottou, Leon & Weinberger, Kilian Q., eds. Advances in neural information processing systems (NIPS 2012), 2012 Lake Tahoe. NIPS, 1097-1105. [NIPS proceedings] (17,181 citations as of 23.11.2017)

[3] http://hci-kdd.org/project/makepatho/

[4] http://hci-kdd.org/2017/10/26/digital-pathology-worlds-fastest-scanner-now-working-graz/

[5] 706.505 Projekt Informationssysteme, 706.800, 706.801 Bachlorarbeit Informatik, 706.804, 706.805, Bachelorarbeit Softwarenetwicklung-Wirtschaft, 706.902 Bachelorarbeit Information and Computer Engineering, 706.501 Master-Projekt, 706.502 Master Project Web and Data Science, and 706.046 AK HCI

https://online.tugraz.at/tug_online/wblvangebot.wbshowlvoffer?ppersonnr=5313

 

 

SP.62 Explainable AI (XAI)
Status
open
Worktype
Bachelor, Master, Ph.D.
Keywords
interactive machine learning
Goals
In the context of our interactive machine learning project iML [1] we have some open work. A huge motivation for this approach are rising legal and privacy aspects, e.g. with the new European General Data Protection Regulation (GDPR and ISO/IEC 27001) entering into force on May, 25, 2018, will make black-box approaches difficult to use in business, because they are not able to explain why a decision has been made, see [2].
Required Knowledge
Interest in Machine Learning and Causality and/or Law
SP.61 Machine Learning for Digital Pathology MAKEpatho
Status
open
Worktype
Bachelor, Master, Ph.D.
Keywords
Machine Learning, Knowledge Extraction, Digital Pathology
Goals
In the context of our large-scale Digital Pathology project MAKEpatho [1] we are currently offering some theses in an exciting, challengenging and brand new environment [2].
For Students of
Students with interest in machine learning/artifical intelligence applied to health informatics
Abstract
There are several work streams available, e.g. in eye-tracking microscope and sensor fusion for several studies to develop a mathematical model of observed cognitions of human experts versus novices, and for detecting patterns of cognition; the work is aimed at the comparison between human learning/decision making and machine learning/prediction and how both can benefit from each other in particular. More details on demand.
Reading
Please check with the students pages for general information [3]
SP.60 Context learning for educational recommender systems
Status
open
Worktype
MSc/PhD thesis
Keywords
automatic machine learning
OESTAT Topics
machine learning
Goals
Our data-driven industry needs a new kind of education to enable students to solve practical problems by application of machine learning/artificial intelligence. However, learning machine learning is not an easy task. Machine Learning today is becoming the most widely and most rapidly growing field in computer science and even to get an top-level overview is a difficult task for a beginner. Even for an enthusiastic student with keen interest in this topic, it is very hard to find where to start and on which topics to concentrate as the field is so immensely broad. This is even harder when the student is interested in applied machine learning, e.g. applied to the health domain - where issues of privacy, data protection, safety and security are no longer nice-to-have topics but a must. However, it is even harder for a teacher to teach a heterogenous group of students with various background knowledge and previous experiences. Consequently, the design of a study plan remains heavily a manual task for the teacher, biased on his personal preferences and background - and less on facts from the current market and industrial demands. As it is simply impossible today to cover each and every subtopic, it makes sense to think on methods for automatic curriculum design, ultimately addressing the current needs and demands of the what is needed by the industry - outside of the ivory tower.
For Students of
Software Engineering, Informatics
Required Knowledge
Interest in Machine Learning, Interest in Human Learning and Teaching
Abstract
This work shall explore concept Web graphs from online sources, e.g. Wikipedia as data sets [2] as recommender systems [3] to bring in line previous knowledge of an individual with required goals of a wished application area (e.g. machine learning for health informatics [1], solving partiuclar problems which can be defined). Ideally, it should be recommended the minimum basic skill set required to solve the problems and giving future outlook for possible specializations (please read the papers below)
Additional Information

A sample curriculum for the course Machine Learning for Health Informatics can be found in [1]

[1] Holzinger, A. 2016. Machine Learning for Health Informatics. In: Holzinger, A. (ed.) Machine Learning for Health Informatics: State-of-the-Art and Future Challenges, Lecture Notes in Artificial Intelligence LNAI 9605, Cham: Springer International Publishing, pp. 1-24, doi:10.1007/978-3-319-50478-0_1

[2] Agrawal, R., Golshan, B. & Papalexakis, E. 2016. Toward Data-Driven Design of Educational Courses: A Feasibility Study.  Journal of Educational Data Mining (JEDM), 8, (1), 1-21.

[3] Calero Valdez, A., Ziefle, M., Verbert, K., Felfernig, A. & Holzinger, A. 2016. Recommender Systems for Health Informatics: State-of-the-Art and Future Perspectives. In: Holzinger, A. (ed.) Machine Learning for Health Informatics: State-of-the-Art and Future Challenges. Cham: Springer International Publishing, pp. 391-414, doi:10.1007/978-3-319-50478-0_20

 

SP.59 ABC-TUGROVIS
Status
open
Worktype
ABC - Agent based Cancer Simulation
Keywords
machine learning, multi-agents
Abstract
Contrary to cellular automata approaches, discrete agent based models are very promising for simulating tumor growth. These can be extended to hybrid systems, also with the biologist-in-the-loop. In this master thesis (which can be the ground work for a PhD) it shall be experimented with state-of-the-art methods and contributed to a larger project on machine learning for tumour growth simulation and visualization to help to contribute towards two goals: supporting cancer resarch and help to reduce animal experiments.
Additional Information

Jeanquartier, F., Jean-Quartier, C., Cemernek, D. & Holzinger, A. 2016. In silico modeling for tumor growth visualization. BMC Systems Biology, 10, (1), 1-15, doi:10.1186/s12918-016-0318-8.

TUGROVIS Project Page

 

SP.58 Scenario-Based Transfer Learning Models
Status
open
Worktype
All levels, work will be adapted accordingly
Keywords
machine learning, transfer learning, agnostic learning
Required Knowledge
Python
Abstract
Humans are very good in transfer learning, i.e. we can solve new problems with regard to previously learned tasks, the classic example is to learn snowboarding from windsurfing, but maybe a better example is language learning: It is much easier to learn Spanish when knowing Italian. Automatic machine learning makes usually no use of such advantages and this is known as the problem of catastrophic forgetting. For two decades now this is an extremely hot topic in the machine learning community [1] and any advances may results in major breaktroughs in this area. In this work the student shall experiment with health related data and formalize some ideas on how to solve concrete transfer learning problems including a benchmarking and evaluation to existing transfer learning approaches.
Additional Information

[1] Thrun, S. 1996. Is learning the n-th thing any easier than learning the first? Advances in neural information processing systems (NIPS), 640-646.
online available

SP.50 Web-based 3D graph visualization and interaction with Canvas and WebGL
Status
closed
Worktype
Bachelor or Master Thesis - the complexity level will be adapted accordingly
Keywords
Graph visualization, Web based Visualization, Canvas, WebGL, X3DOM, Fast & Fluid Web Applications
OESTAT Topics
1108
Goals
This work will be part of a larger research project called iKNOdis.net (interactive knowledge discovery in networks), whose goal is object recognition and classification via graph based data mining on natural images. In order to visualize the results of every phase in this process as well as being able to show the functionality of graph algorithms step-by-step, we need strong 3D visualization capabilities.
Traditionally the browser has been unable to accomplish complex visualization tasks, however, with the advent of WebGL and the HTML canvas element some years ago, the field has gained traction and a cornucopia of visualization libraries has been developed ever since, enabling activities ranging from simple 2D-chart visualization of datasets to web based game programming. As part of our ongoing research, students have already undertaken to implement 3D graph visualization utilizing the three.js and scene.js libraries. Although successful in their task, there were still a lot of performance problems and a lack of interaction possibilities. Therefore in this work, you will evaluate different routes to utilize Canvas & WebGL, building upon the work already done, but extending your insights to writing low-level WebGL yourself as well as other alternatives such as X3DOM (a way to write WebGL like HTML, in loose terms..). The emphasis of this project lies on performance, as we will need to visualize several thousand nodes (and about an order of magnitude more edges) in near-realtime. Upon comparing the alternatives, you will pick a winner and implement a proof-of-concept; as our framework for extracting graphs out of images inside the browser already exists, the results of your efforts could have a direct impact on our real-world application!
For Students of
Software Engineering, Informatics (Mathematics, Telematics)
Required Knowledge
Programming skills, knowledge about web based development, interest in modern browser based visualization and interaction, 3D web apps
SP.49 Canvas Alternatives for Web-based 3D graph visualization and interaction
Status
closed
Worktype
Bachelor or Master Thesis - the complexity level will be adapted accordingly
Keywords
Graph visualization, Web based Visualization, SVG, CSS 3D, Fast & Fluid Web Applications
OESTAT Topics
1108
Goals
This work will be part of a larger research project called iKNOdis.net (interactive knowledge discovery in networks), whose goal is object recognition and classification via graph based data mining on natural images. In order to visualize the results of every phase in this process as well as being able to show the functionality of graph algorithms step-by-step, we need strong 3D visualization capabilities.
In recent years such capabilities have been available through the use of WebGL and the HTML Canvas element, however this approach has several drawbacks: The canvas itself is isolated from the rest of the browser DOM and therefore elements within it cannot be styled via CSS attributes or subjected to CSS event handling (hover, active, visited, etc.). Moreover, since elements within canvas are actually no elements at all (just pixels computed in a certain fashion), no Javascript event handling can be applied to them on an element (object) basis. This makes any interaction with a scene rendered in canvas very complicated: Canvas based libraries use raytracing to identify which elements might be 'under' a mouse pointer and then recompute the scene in some way - but the programmer cannot easily influence / extend the library's functionality. Therefore, this work consists of the research and evaluation of alternatives to canvas based 3dimensional visualization and the implementation of a proof-of-concept. Two logical candidates are SVG and CSS-3D; whereas SVG uses it's own markup syntax, CSS-3D is purely browser based and was hitherto considered too slow for any serious work. Very recently however (as of 2014), libraries like Facebook's react.js and the infamous famous.js present promising options even on mobile devices (see the reading section). If you are interested in modern client side web programming and HTML5 for mobile devices, this is your project!
For Students of
Software Engineering, Informatics (Mathematics, Telematics)
Required Knowledge
Programming skills, knowledge about web based development, interest in modern browser based visualization and interaction, 3D web apps
SP.48 WebGL for non-graphical Parallel Computing
Status
closed
Worktype
Bachelor or Master Thesis - the complexity level will be adapted accordingly
Keywords
Parallel Computing, WebGL, Actor Model, Belief Propagation, Bayesian Networks, Markov Random Fields
OESTAT Topics
1108
Goals
This work will be part of a larger research project called iKNOdis.net (interactive knowledge discovery in networks), whose goal is object recognition and classification via graph based data mining on natural images. As a first step to this goal, we extract graphs out of images where the nodes represent granular regions which we treat as semantic atoms (e.g. a cell in a skin tissue sample). One way to analyse the resulting graph structure is by belief propagation in Markov Random Fields (MRF, which are similar to Bayesian networks but undirected and potentially cyclic). In this approach, every node receives an initial belief about itself ('i am a cancer cell') based on which it forms opinions about its neighboring nodes (e.g. 'i am a cancer cell - you are my direct neighbor and not on my gradient - so you must be a cancer cell too'). In every step of the process, each node then propagates its opinions to its neighbors, which incorporate all incoming messages into their new belief about themselves ('i am 76% sure i am not a cancer cell'). This process is repeated until the system reaches a coherent state of beliefs, or is aborted. Such kinds of algorithms are very suited for Actor frameworks (such as Akka for java), which allows every 'actor' (=node) to be programmed as a singular object with message sending and receiving capabilities; the framework takes care of the rest. As the iKNOdis team strives to build a web based research platform and all computations will be done in the browser, our choices are limited: Either use Javascript - and try to imitate a framework like Akka - or utilize the computer's GPU, whose architecture was designed from the ground up for massive parallel computing. In this work, you will explore the possibility of using normal GPU Shaders to simulate a belief propagation network as described above and implement a proof-of-concept; as we already have the browser-side code necessary to extract graphs out of images, your work will be immediately applicable and could help setting new standards in automatic image classification (skin cancer detection).


This is a great opportunity for hardcore programmers; further details (including payment) to be elaborated in a personal meeting!
For Students of
Informatics, Mathematics, Telematics, Software Engineering
Required Knowledge
Basics in data structures and algorithms as well as programming skills necessary, interest in parallel computing inside WebBrowsers, WebGL architecture and language
SP.47 Content Based Image Retrievel - DB evaluation and proof of concept
Status
closed
Worktype
Bachelor or Master Thesis - the complexity level will be adapted accordingly
Keywords
Object Representation, Object Database, Fast Image Retrieval, Sorting / Indexing on complex objects
OESTAT Topics
1108
Goals
This work will be part of a larger research project called iKNOdis.net (interactive knowledge discovery in networks), whose goal is object recognition and classification via graph based data mining on natural images. One method of doing this is to derive global image understanding from already known local primitives - objects - contained in those images. However, doing this efficiently requires 1) a feasible (=small) and unique representation of those objects as well as a means of retrieving their semantic meaning efficiently from a database. To facilitate such an approach, CBIR systems have been proposed that not only store and organize / index object representations but also take object content as a search query. The tasks for this project will comprise the research & evaluation of existing content based database systems, choosing a particular technology suited for the job at hand, and the implementation of a proof-of-concept. As no (open source) standard for CBIR systems has evolved as of date, one possible solution could consist of 'enhancing' an already existing database system - MongoDB, Redis, etc. - to handle this problem by computing object similarities internally. This project is also suitable for a team of two, if the second team member covers the 'Object Modeling' project listed on this page, which overlaps strongly with this work.
For Students of
Informatics, Mathematics, Telematics, Software Engineering
Required Knowledge
Basics in data structures and algorithms as well as programming skills necessary, interest in modern database architectures (NoSQL, GraphDatabases), nearest neighbor search
SP.46 Object Modeling - Evaluate and implement unique and invariant object representations
Status
closed
Worktype
Bachelor or Master Thesis - the complexity level will be adapted accordingly
Keywords
Object Recognition, Object Modeling & Representation, Object Database, Datastructure Matching, Image Retrieval
OESTAT Topics
1108
Goals
This work will be part of a larger research project called iKNOdis.net (interactive knowledge discovery in networks), whose goal is object recognition and classification via graph based data mining on natural images. To this end, it is necessary to identify semantically granular parts of images - objects - in order to be able to derive global image understanding from local phenomena and their relations. This can only be done if objects located in an image can be efficiently matched to templates stored in a database, which in turn requires unique and stable (= invariant to rotation, scale, ..) representations of the objects in question. Subject of this Bachelor / Master thesis will be the research and evaluation of different object representation methods (feature vectors, glyph representation, mixed distribution models etc.). Upon identifying a suitable method, the practical part will consist of the development and implementation of a proof of concept. As the iKNOdis team strives to provide a web based research platform, the software will be implemented in JavaScript (Typescript). This problem can also be tackled in a team of two, if the second team member covers the 'Content Based Image Retrieval' project listed on this page.
For Students of
Informatics, Mathematics, Telematics, Software Engineering
Required Knowledge
Basics in data structures and algorithms as well as programming skills necessary, interest in feature extraction & object representation
SP.41 The Web as biomedical laboratory – from Wet-lab to Web-lab
Status
closed
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
Web, Mobile computing, e-Science, computational biology, big data, open data
OESTAT Topics
1108, 3927
Goals
Wet lab experiments in the biomedical domain are very expensive and time consuming and locally bounded. Due to the advances in Web Sciences, along with the huge availability of open source software, open access publications and open biomedical data, there is big chance to use the Web as a useful multi-disciplinary lab environment: from wet-lab to web-lab. A grand challenge is to research on how to best leverage the potential of such experimental spaces, i.e. what are constraints for web-lab experiments? What are acceptable measures? What about robustness and repeatability? What are the limitations in terms of computational power? What are the opportunities for experimental world-wide collaboration? There are a lot of research avenues open for design, development and experimental work on this topic of e-Science.
For Students of
Informatics, Mathematics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in experimental work and e-Science
Reading
Marx, V. (2012). My data are your data. Nature Biotechnology, 30, (6), 509-511. <link>

 
SP.30 BIOMETEX - Biomedical Knowledge Discovery from MEDLINE applying information theoretical approaches
Status
closed
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
SP.30 Biomedical Text Mining, Knowledge Discovery, large data sets, pointwise mutual information, information theory, entropy
OESTAT Topics
1108, 3927
Goals
The goal of this project is to extent previous approaches and test advanced methods of information theory on the basis of real-world examples from natural language processing. The results may contribute towards revealing specific hidden knowledge within non-standardised text, thereby exposing relationships which can provide important additional information for medical experts (more information will be provided in a personal meeting).
For Students of
Informatics, Mathematics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in text mining, information retrieval and knowledge discovery methods
Reading
[1] MEDLINE - Medical Literature Analysis and Retrieval System Online
[2] Davenport, T. H. & Glaser, J. (2002) Just-in-time delivery comes to knowledge management. Harvard Business Review, 80, 7, 107-111.
[3] Holzinger, A.; Simonic, K.-M.; Yildirim, P. (2012) Disease-disease relationships for rheumatic diseases Web-based biomedical textmining and knowledge discovery to assist medical decision making. In: IEEE COMPSAC, 36th Annual International Computer Software and Applications Conference [Preprint of the paper for download]
Additional Information

Biomedical Knowledge DiscoveryBiomedical discoveries are documented mostly in scientific articles. Web-based collections of such biomedical articles contain an exponentially increasing amount of text data. For example the MEDLINE – Medical Literature Analysis and Retrieval System Online [1] – is the National Library of Medicine’s (NLM) premier bibliographic database that contains 19+ million references to journal articles in life science and biomedicine. The records of this database are indexed by NLM Medical Subject Headings (MeSH). As a matter of fact, yet 10 years ago, a typical medical practitioner had to stay up-to-date of more than 10k diseases and syndromes, more than 3k medications and more than 1k lab tests [2]. So it is easy to understand that no human medical expert can manually transfer all articles dealing with relevant information into knowledge; but what the experts can is asking questions, stating hypotheses. Consequently, there is an urgent need for methods and tools to enable the expert to discover relevant hidden knowledge in those massive data sets. For this purpose intelligent interactive text mining methods can be applied to assist the reader, e.g. to look for similarities or anomalies within these large volumes of these non standardized textual data. Statistical models can be used to evaluate the significance of the relationship between entities such as disease names, drug names, and keywords in titles, abstracts or within the entire publication [3].

SP.29 IVISMEDOC – Interactive knowledge visualization of medical documents using probabilistic topic models
Status
open
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
SP.29 Biomedical Knowledge Discovery, Topic modelling, visualization
OESTAT Topics
1108, 3927
Goals
The goal of this project is to apply topic modelling algorithmus to collections of medical documents aiming at structuring to help professional readers to identify the most interesting parts of a document. Moreover, the hidden topic proportions implicitly connect each document to the other documents (e.g. by considering a distance measure between topic proportions). Some questions include: How can we best display these connections? What is an effective interface to the whole corpus and its inferred topic structure? These are user interface questions, which are of tremendous importance in topic modeling.
For Students of
Informatics, Mathematics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in solving real-world problems, real-world data, interest in knowledge discovery
Reading
Blei, D. M. (2012) Probabilistic topic models. Communications of the ACM, 55, 4, 77-84
Additional Information

Collective data in medical documents is vastly increasing, making it more and more difficult to discover relevant knowledge. One possible approach of harnessing this data deluge is in the application of topic modeling. Topic modeling algorithms (e.g. Latent Dirichlet Algorithm, LDA) can be applied to large collections of documents, e.g. to find previously unknown patterns.  A promising future direction for topic modeling is to develop new methods of interacting with and visualizing topics and corpora. Topic models provide new exploratory structures in large collections, the question remains on how can we best exploit that structure to aid interactive knowledge discovery.
A big problem is how to visualize the topics. Generally, topics are visualized just by listing the most frequent keywords of a document; however, new ways of labeling such topics may be more effective. A further problem is how to best display a document with a topic model. At the document level, topic models provide potentially useful information about the structure of the document.
Topic modeling algorithms show much promise for uncovering meaningful thematic structure in large collections of documents. But making this structure useful requires careful attention to information visualization and the corresponding user interfaces. Consequently, the goal of this work is to experiment with medical documents on possible solutions and to experimentally test the solutions in the medical area.

SP.23 EMOMES - Computational Emotion Detection and Measuring
Status
closed
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
SP.23 Emotion, Emotion detection, Physiological measurement,
OESTAT Topics
1108, 3927
Goals
The goal of this work is to experiment with approaches on how to measure emotions and how to include this into useful applications for our daily work, e.g. in the application of medical information systems. The central hypothesis is: Measured emotional values can be applied for computational decision support.The challenge of this work is to develop an unobtrusive, easy to use and low cost emotion measuring application.
For Students of
Informatics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in HCI, measurement
Abstract
Emotions are regarded as important mental and physiological states influencing perception and cognition. Consequently, emotion has enourmous influence on cognitive performance during all types of human activities. Example applications include decision making, recommender systems, learning etc. Therefore, the concept of emotion is a topic of enourmous interest in Human-Computer Interaction (HCI) for some time. Popular examples include stress detection or affective computing. The challenge of this work is to develop an unobtrusive, easy to use and low cost emotion measuring application.
Reading
[1] Dolan, R. J. (2002) Emotion, Cognition, and Behavior. Science, 298, 5596, 1191-1194.
Gong, X., Yang, Y., Lin, J. H. & Li, T. R. (2011) Expression Detection Based on a Novel Emotion Recognition Method. International Journal of Computational Intelligence Systems, 4, 1, 44-53.
Lopatovska, I. & Arapakis, I. (2011) Theories, methods and current research on emotions in library and information science, information retrieval and human–computer interaction. Information Processing & Management, 47, 4, 575-592.
Additional Information

Emotion Kognition VerhaltenEmotions are regarded as important mental and physiological states influencing perception and cognition, thus behaviour [1]. Consequently, emotion has enourmous influence on cognitive performance during all types of human activities. Example applications include decision making, recommender systems, learning etc. Therefore, the concept of emotion is a topic of enourmous interest in Human-Computer Interaction (HCI) for some time. Popular examples include stress detection or affective computing.

SP.22 TEDAVISMED - Testing Data Visualization Methods in the medical domain
Status
closed
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
SP.22, Visualization, Visual Analytics, Interactive Visualization, Data Visualization
OESTAT Topics
1108, 3927
Goals
There are so many sophisticated visualization methods available, e.g. Parallel Coordinates, RadViz, Radarplots, heatmaps, Glyphs, Chernoff faces, etc. etc., however, they are rarely applied in business enterprise hospital information systems. The question remains open: why? What is the reason that most of such elaborated visualization methods are used predominantly in academic environments? This work shall develop a test bed setting, which can be used for pre-test post-test experimental performance evaluation of different tasks supported with different visualization methods. The central hypothesis is “Method X is superior than method Y in solving problem Z”. The outcome is a systematic analysis of advantages/disadvantages of each method, benchmarked against a standard real-world data set.
For Students of
Informatics, Mathematics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in Data Visualization, high dimensional complex medical data, testing in real-world environments
Additional Information

According to Johnson (2004) [1] developers of visualization technologies do not spend enough (or indeed any) time endeavoring to understand the underlying scientific aspects they are trying to represent, just as application scientists sometimes create crude visualizations without understanding the algorithms and the science of the visualization, respectively. It is necessary to understanding the underlying science, engineering, and medical applications. There is no substitute for working together with end users to create better techniques and tools for solving challenging scientific problems. Issues of cognition, perception and reasoning are of great importance and tasks such as discovering patterns of change in the data will involve not only data visualization, but also how the data is changing. Such high-level discoveries can be used by the domain analyst to form, confirm, or refute a set hypothesis, expand or correct mental models, and provide confidence in decision making processes – which is still the core research area in biomedical informatics [2].

SP.21 GAMSMEDLE - Game-based Medical Science Learning
Status
open
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
Game based Learning; Science learning; Biomedical education; Science education
OESTAT Topics
1108, 3927
Goals
There is a huge potential in game based learning approaches. Within a large project, we offer various opportunities on the design, development and experimental evaluation of game based learning approaches in different contextual environments and different medical domains.
For Students of
Informatics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in Game-based Learning and Medical Education
Reading
[1] Strauss, S. (2012) Gamers outdo computers at matching up disease genes: Computer game crowdsources DNA sequences alignment across different species. [Nature, News, March, 12]

[2] Foldit [Start page]

[3] Foldit [Nature Video]

[4] Cooper, S., Khatib, F., Treuille, A., Barbero, J., Lee, J., Beenen, M., Leaver-Fay, A., Baker, D., Popovic, Z. & players, F. (2010) Predicting protein structures with a multiplayer online game. Nature, 466, 7307, 756-760.

[5] Kawrykow, A., Roumanis, G., Kam, A., Kwak, D., Leung, C., Wu, C., Zarour, E., Sarmenta, L., Blanchette, M., Waldispühl, J. & Phylo, p. (2012) Phylo: A Citizen Science Approach for Improving Multiple Sequence Alignment. Plos One, 7, 3, e31362.

[6] Ebner, M. & Holzinger, A. (2007) Successful Implementation of User-Centered Game Based Learning in Higher Education – an Example from Civil Engineering. Computers & Education, 49, 3, 873-890.

[7] Mayo, M. J. (2009) Video Games: A Route to Large-Scale STEM Education? Science, 323, 5910, 79-82.

[8] Holzinger, A., Nischelwitzer, A., Friedl, S. & Hu, B. (2010) Towards life long learning: three models for ubiquitous applications. Wireless Communications and Mobile Computing, 10, 10, 1350-1365.

[9] The concept of Life Long Learning [Video]
Additional Information

Science Game based LearningAn excellent example of a science game is Foldit (Strauss, 2012) [1], [2], [3]: Protein structures are important for many purposes in medicine and the life sciences and are specifically studied in molecular biology and bioinformatics respectively. Successful identification of structural topologies of proteins enable experts to study and understand protein-protein interactions (PPI), which may lead to the creation of new proteins along with advancements in the treatment of diseases and many other biomedical problems (Cooper et al., 2010) [4]. Foldit is an online puzzle game with the goal of folding the structure of selected proteins to the best of the player’s ability, using various tools provided within the game. The highest scoring solutions are analyzed by researchers, who determine whether and to what extent there is a native structural configuration (or native state) that can be applied to the relevant proteins back in the real-world. This is also a good example for crowdsourcing . A further outstanding example is Phylo, a human-based computing framework applying crowdsourcing techniques to solve the Multiple Sequence Alignment (MSA) problem; the key idea of this game is to convert the MSA problem into a casual game that can be played by ordinary web users with a minimal previous knowledge of the biological context (Kawrykow et al., 2012) [5]. The power of games for educational purposes has been proved in several areas (Ebner & Holzinger, 2007) [6], (Mayo, 2009) [7] and are also useful for inclusion in Life Long Learning scenarios (Holzinger et al., 2010) [8], [9].

SP.20 iMOFECU - Interactive Mobile Fever Curve - Time oriented data visualization on high-res tablet computers
Status
closed
Worktype
all types of student work/theses possible - the complexity level will be adapted accordingly
Keywords
time oriented data, visualization, longitudinal data, mobile device, iPad3, tablet, bedside computing
OESTAT Topics
1108, 3927
Goals
The goal of this pioneer project is to design, develop and experiment with various prototypes of electronic “fever curves” (German: Fieberkurve) the bedside patient chart, using most advanced high-resolution tablet computers. The central hypothesis is that the information density of a paper chart can be brought reasonable to a high-resolution tablet computer (specific and detailed information will be given on demand)
For Students of
Software Engineering
Required Knowledge
Interest in design, developing and experimenting with high resoution touch tablet computers
Additional Information

Fieberkurve - bed side patient chartThe bed side paper patient chart (sloppy called “fever curve”) is of invaluable importance for the daily work of the clinician. It is an important summary of medical relevant data and a substantial part of the patient
documentation, necessary for the physician and for everyone concerned with patient care (nurses, therapists, shared care etc.). However, to date, no convincing, clinical useable and useful electronic solution is available. Previous attempts mostly failed on the insufficient display resolution of the devices. Consequently, the recent Super AMOLED capacitive touchscreens with 300+ ppi pixel-density could possibly be a first chance to make such a chart of benefit for the clinician (previous work available – more information on demand).

SP.18 AAEKNO - Applying Approximate Entropy to discover knowledge out of complex noisy medical time-series data
Status
closed
Worktype
SP.18 All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
Biomedical Informatics, Approximate Entropy, Time-Series data
OESTAT Topics
1108, 3927
Goals
The goal of this project is to experiment with approximate entropy methods on time series data with the challenge to design, develop and evaluate a tool comprising a special end-user-centered interface, so that a medical expert having little computer literacy can have a look on complex noisy time series data and may discover relevant knowledge out of these data (details will be given in a personal meeting).
For Students of
Informatics, Telematics, Biomedical Engineering, Software Engineering, Mathematics
Required Knowledge
Interest in data mining, computational science, especially statistical learning and work on "real-world" problems with non-computer literate end-users
Reading
[1] Simonic, K. M., Holzinger, A., Bloice, M. & Hermann, J. (2011) Optimizing Long-Term Treatment of Rheumatoid Arthritis with Systematic Documentation. Proceedings of Pervasive Health - 5th International Conference on Pervasive Computing Technologies for Healthcare. Dublin, IEEE, 550-554.
[2] Pincus, S. M. (1991) Approximate Entropy as a measure of system complexity. Proceedings of the National Academy of Sciences of the United States of America, 88, 6, 2297-2301.
(more information on demand)
Additional Information

We can see time series data as a set of data points representing observations made in time. If the observations are made subsequently in time we speak of an equally spaced time series. In medical practice often it is not possible to achieve equally spaced time series (Simonic et al., 2011), [1]. The main problem involved is, that standard methods may model artefacts. By application of approximate entropy (ApEn) we are able to classify complex systems, which may include both deterministic chaotic and stochastic processes (Pincus, 1991), [2].

SP.17 PERSTATLE - Performance Testing of selected Statistical Learning Methods on a real-world sample data set
Status
open
Worktype
SP.17 All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
Knowledge Discovery, Data Mining, Statistical learning
OESTAT Topics
1108, 3927
Goals
The aim of this work is to experiment with selected standard and non-standard methods for learning from data on a real world sample data set in order to systematically determine various advantages and disadvantages.
For Students of
Informatics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in Knowledge Discovery, Data Mining and statistical learning and an interest in working with "real-world" medical data
Reading
[1] Hastie, T., Tibshirani, R. & Friedman, J. (2009) The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Second Edition. New York, Springer. [Online available]
Additional Information

statistical learningComputational problems in biology, medicine and life sciences create masses of data and the primary goal is in sensemaking, i.e. to extract relevant patterns and trends, and to gain out knowledge of this data. Hastie, Tibshirani and Friedman call this “learning from data” (Hastie, Tibshirani & Friedman, 2009) [1]. The learning problems which they consider are generally categorized in supervised learning (e.g. Bayesian statistics, nearest neighbor algorithm, support vector machines, etc.), or unsupervised learning (e.g. hierarchical clustering, principal component analysis, neural network models, etc.). In supervised learning, the goal is to predict the value of an outcome measure based on a number of input measures (set of training data), whereas in unsupervised learning, there is no outcome measure, and the goal is to describe the associations and patterns among a set of input measures (typically vectors).

SP.16 KNODWAT - KNOwledge Discovery With Advanced Techniques
Status
closed
Worktype
Master Thesis
Keywords
Knowledge Discovery, Data Mining, Machine learning, Human-Computer Interaction, Data Visualization, Data Analytics
OESTAT Topics
1108, 1161
Goals
The field of knowledge discovery and its applications are vast and widely studied, however, for novice researchers who are unexperienced in the use of even relatively popular knowledge discovery methods, it can be difficult to find useful, practically oriented tools and methods, which do not require extensive previous knowledge of the subject. It can be even harder for inexperienced users, to actually apply such methods to generate useful information. The idea of KNODWAT is to design, develop and test a a web-based, highly extensible scientific framework application for \"playing\" with knowledge discovery methods.
For Students of
Finished Master Thesis by Mario Zupan, thesis defence on November, 29, 2012 (passed with distinction)
SP.15 EUPMED - End-User Programming in Medicine
Status
closed
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
SP.15, End User Programming, mashups, health care, usability, human-computer interaction
Goals
The goal of this project is to design, develop and evaluate a test bed which shall serve two purposes: to test the concept of mashups (hybrid web applications) along with testing the concept of end-user programming within a clinical setting. The research focus is on finding harmonies/differences between the thinking of programmers and non-programmers, in order to optimize information interaction in End-User programming. The central hypothesis is, that user-centred mashups improve the performance of end-users in accomplishing complex tasks.
For Students of
Software Engineering Business (924) o.a.
Required Knowledge
Interest in working in a real-life domain, interest in medical workflows
Reading
[1] Adams, S. S. (2008) The future of end user programming? Companion of the 30th international conference on Software engineering. Leipzig, Germany, ACM, 887-888.
[2] Auinger, A., Ebner, M., Nedbal, D. & Holzinger, A. (2009) Mixing Content and Endless Collaboration – MashUps: Towards Future Personal Learning Environments. In: Stephanidis, C. (Ed.) Universal Access in Human-Computer Interaction HCI, Part III: Applications and Services, HCI International 2009, Lecture Notes in Computer Science (LNCS 5616). Berlin, Heidelberg, New York, Springer, 14-23.
[3] Aghaee, S. & Pautasso, C. (2012) End-User Programming for Web Mashups. In: Harth, A. & Koch, N. (Eds.) Current Trends in Web Engineering. LNCS 7059. Berlin, Heidelberg, Springer, 347-351.

Cao, J., Riche, Y., Wiedenbeck, S., Burnett, M. & Grigoreanu, V. (2010). End-User Mashup Programming: Through the Design Lens. CHI 2010: 28th Annual Conference on Human Factors in Computing Systems, New York, Association of Computing Machinery, 1009-1018.
Holzinger, A., Mayr, S., Slany, W. & Debevc, M. (2010). The influence of AJAX on Web Usability. ICE-B 2010 - ICETE The International Joint Conference on e-Business and Telecommunications, Athens (Greece), INSTIC IEEE, 124-127.
Wong, J. & Hong, J. I. (2007). Making Mashups with Marmite: Towards End-User Programming for the Web. ACM Conference on Human Factors in Computing Systems, Vol 1 and 2, New York, Association of Computing Machinery, 1435-1444.
Additional Information

End User Programing EUP„The Web is filled with millions of customized applications, most created by end users themselves (Adams, 2008), [1]”. Hybrid web applications, so called mash-ups, are promising web service concepts, allowing end users with very low computer literacy (end users with minimal or no programming experience) to tailor their software applications exactly to their tasks, their needs and their environment (Auinger et al., 2009), [2]. Whilst there are some examples from diverse areas, research in the area of clinical medicine and health care is highly necessary (Aghaee & Pautasso, 2012) [3].

SP.11 TWIPUMED - Twitter analyzator to enhance public medical predictions
Status
closed
Worktype
All types of student work/theses possible (e.g. 706.403, 706.404, 706.116, 706.119 etc.) - the complexity level will be adapted accordingly
Keywords
SP.11 Network Theory, sensemaking, time series data, social network, twitter, app
OESTAT Topics
1108, 1109, 1150, 1157
Goals
The goal of this work is to design, develop and evaluate a tool for end-user centred medical analysis of social media streams. The central hypothesis is, that such a tool brings benefits and supports research on public health (detailed information will be given in a personal meeting).
For Students of
Informatics, Mathematics, Telematics, Biomedical Engineering, Software Engineering
Required Knowledge
Interest in network theory, graph theory; good mathematical background
Reading
[1] Savage, N. (2011) Twitter as medium and message. Communications of the ACM, 54, 3, 18-20.
[2] Hawn, C. (2009) Take two aspirin and tweet me in the morning: how Twitter, Facebook, and other social media are reshaping health care. Health affairs, 28, 2, 361-368.
[3] Cheng, Z., Caverlee, J. & Lee, K. (2010). You are where you tweet: a content-based approach to geo-locating twitter users. ACM, 759-768.
Additional Information

Network scienceWeb Sciences encompass the study of social networks, for example to gain knowledge out of big social data. Social media allow getting an impression of the thoughts and activities of the society at large. Consequently, Twitter data may help to answer questions or to support hypotheses, which are otherwise hard to approach, because a manual polling of large groups of people is too time consuming, too expensive or just impossible (Savage, 2011) [1], (Hawn, 2009) [2]. To analyze such data is of high interest for public health measures. A typical example is the correlation between mined Twitter messages with actual influenza rates. There are many research & development possibilities of public health apps for Twitter, e.g. tracking illnesses over times. We call this syndromes surveillance, i.e. to monitor clinical syndromes that may have significant impact on public health, impacts medical resource allocation, health policy and education. Moreover, there are possibilities to measure behavioral risk factors, localizing illnesses by geographic region, and/or analyzing symptoms and medication usages (Cheng, Caverlee & Lee, 2010) [3]. Technical approaches include probabilistic topic models, e.g. Latent Dirichlet allocation (LDA).

SP.9 PROMOHEALTH - Design and development of a web application for mobile devices promoting a healthy lifestyle
Status
closed
Worktype
finished Master Thesis F 067 484 Individuelles Masterstudium; Biotechnologie (Biotechnology)
Keywords
SP.09 Health Care, Mobile, Wellness Applications, Health Data Visualization, Usability Research
OESTAT Topics
1108, 3927 Medizinische Informatik
For Students of
finished Master Thesis by Stefan DORNER, exam on 01.04.2011
Additional Information

Health Trend prediction analysisThe decreasing general state of health caused by lack of health consciousness, sedentary lifestyles and demographic changes will have dramatic effects on our health care system in years to come. An unhealthy lifestyle will lead to an increase in chronic diseases and eventually to increased costs in health care. A preventive measure against such a development can be to reinforce health-awareness through the use of mobile applications supporting self-observation and behavior change. The aim of this project was the design and development of a mobile web application that assists people in changing their lifestyles by providing the means to manage their wellness related activities and health risks. The application not merely offers the means for wellness management but also attempts to create high motivation through the adaption of design goals created especially for supporting behavior change. A user study on the final prototype including a questionnaire showed very good usability ratings with a SUS score of 83.75. The majority of our respondents stated that the functions offered by the system could be useful for them and they could image that using the application might motivate them to lead a healthier life. The goal-reward system and the summarizing feedback page were the most popular features among our test users. Usability issues discovered in the study included button size and spacing as well as the system reaction on tapping events. Neither age nor previous experience with computers or smartphones showed a significant influence on the users’ perceived usability and on their motivation to use the application.

SP.8 COGLUME - Computational Mobile Glucose Management for Decision Support
Status
closed
Worktype
finished Master Thesis F 066 924 Masterstudium; Softwareentwicklung-Wirtschaft (Software Development)
Keywords
SP.08 Clinical software; decision support; diabetes; user-centred development; Android; mobile computing; medicine, tablet; touch; glucose management
OESTAT Topics
1108, 3927
For Students of
finished Master Thesis by Bernhard HOELL, exam on 27.10.2011
Required Knowledge
Interest in Mobile Computing, Biomedical Informatics, User-Interface design, Android Development
Additional Information

Medical Touch tablet pcThis master thesis deals with the design and development process of a mobile Android application to support the inpatient glucose management of patients with diabetes at the University Hospital Graz in order to optimize the current paper based glucose management. An integrated decision support service for insulin dosing should provide additional security and support for clinicians. The master thesis was carried out in the course of the EU-project REACTION at the Joanneum Research institute HEALTH – Institute for Biomedicine and Health Sciences – in Graz. The thesis is generally divided into two parts. The first part deals with an extensive requirements analysis, in order to get an imagination of the design of the application’s user interface, as well as to understand clinical workflow patterns. The design phase followed a user-centered design approach, which means that the end-users have been involved in every step of the design process. In the second part, the achievements of the requirements analysis were used to set up the implementation of the inpatient glucose management system. Due to maintainability and expandability it was decided to distinguish between a frontend application for user interactions and a platform independent backend application, which contains the business logic for the decision support, as well as the data storage and interfaces to the hospital information system. The exchange of data between the backend and the frontend is done via encrypted web services to provide data security. This master thesis primarily deals with the development of the frontend apllication and should illustrate collected experiences during the design and the development process. It should demonstrate the requirements and challenges of implementing safety-critical medical software and should show how the end user can be involved in the engineering process.

“The most rewarding research is the one that delights the thinker and at the same time is beneficial to humankind” (Christian Doppler, 1803‐1853), consequently, we are devoted to our guiding principle: “Science is to test crazy ideas – Engineering is to put these ideas into Business” (A.Holzinger, 2011, Successful Management of R&D).

Holzinger-Group-Research-Approach